Article pubs.acs.org/est
Online Monitoring of Molecular Processes in a Plasma Air Purifying System Stefan Schmid, Lukas Meier, Christian Berchtold, and Renato Zenobi* ETH Zurich, Department of Chemistry and Applied Biosciences, CH-8093 Zürich Switzerland S Supporting Information *
ABSTRACT: Plasma air purifying systems present an interesting alternative to filters for purifying air. In this study, molecular processes in a commercially available ac driven plasma air purifier were studied in detail. This air purifier is supposed to reduce all air contaminants to small nontoxic molecules (e.g., H2O and CO2). However, degradation mechanisms are not yet fully understood. In this study, we investigated the exhaust of the plasma air purifier to determine which degradation products are formed. An interface was designed and constructed to allow the direct coupling of the plasma air purifier’s exhaust to a mass spectrometer. The compounds studied, primary and secondary amines, were introduced at a concentration of 1 ppmV. Contrary to our expectations, polymerization instead of degradation was observed. The higher the ac voltage applied (max. 9.0 kV) to the plasma air purifier, the higher the mass of the oligomer distribution. Side chain oxidation products as well as oligomers could be observed for all compounds tested. Starting with amines of low mass (m/z < 200), compounds of molecular masses above 1000 Da were observed in the plasma air purifier. Detailed analysis of the observed mass spectra as well as experiments with deuterated dibutylamine helped to unravel the mechanism taking place in the plasma air purifier. Nitrate anions generated in the plasma air purifier (presumably from N2) are proposed to form ionic clusters with protonated amines.
■
INTRODUCTION Purifiers based on nonthermal plasma are a novel alternative to filters for cleaning air. These systems should be able to degrade any air contaminants to nontoxic compounds, while consuming only limited power (50 mW to 17 W, information from suppliers). In recent years, air cleaning systems have become of great interest to both the general public and the economy. Buildings with ever better thermal insulation, humidity, and temperature control result in a smaller exchange rate with the outdoor air and thus lead to poorer indoor air quality.1−6 Many infectious diseases are caused by airborne pathogens, the large majority of which is spread in indoor environments.7 This takes a big toll on people and on the economy,8 especially because many people spend 80% and more of their time indoors.9−11 Indoor air pollution is among the top five environmental risks to public health, and it is suggested to ventilate buildings and rooms several times a day with clean outdoor air. However, this is not always desirable or possible, due to weather conditions or due to energy saving issues. Therefore, air cleaning systems have become increasingly important. From 1998 to 2007, 51 patents for plasma air purifiers (PAPs) were filed, whereas in the last 3 years close to 100 patents for PAPs were applied for. © 2012 American Chemical Society
This shows the growing general interest in this technology. Several designs of PAPs have been reported; most of them are using a heterogeneous catalyst and relatively high power for achieving good degradation efficiencies (from 70 to 100%).12−18 Until now, there is no standardized method to characterize the degradation efficiency of such plasma-based air purification systems because mechanisms of the reactions and the reaction pathways are still not fully understood. One method to specify the degradation efficiency of a PAP is to study the fate of contaminants such as volatile organic compounds by using adsorption tubes. Although the overall cleaning efficiency can easily be evaluated using adsorption tubes, it is difficult to detect the degradation products generated. Some reactive products formed may not be adsorbed at all, whereas others (e.g., ozone and radicals) 19 could react with the adsorption material in the tubes, which would therefore render subsequent desorption of such compounds impossible. Received: Revised: Accepted: Published: 4067
December 1, 2011 February 9, 2012 March 15, 2012 March 15, 2012 dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
Figure 1. Schematic design of the used system for testing the PAP. A, sample introduction accomplished using a supersonic ionization like spray source; B, PAP consisting of 4 copperplates and 4 plates made of poly-(methyl-methacrylate) as spacers; C, Self-designed Interface, which consist of a heating coil (1), the miniFAPA ion source for secondary ionization (2), an aluminum tube (3), a connection for a vacuum pump (4) and a 150 mm elongated MS cone, whose tip is placed 10 mm in front of the miniFAPA ion source; D, the LCQ-MS for mass analysis.
the system was generated by a fan which sustained a constant air flow of 320 L min−1. The PAP (B) is incorporated into the self-made ventilation system. The diameter of the PAP is 160 mm, its length is 180 mm, and its overall weight is 10.5 kg. The plasma electrodes inside the PAP are made of nickel, lead, and copper, and are supported by polyvinyl chloride and poly(methyl-methacrylate) holders. The plasma in the PAP is generated using air, no auxiliary gas is needed. A high ac voltage is applied to the copper plates with a frequency of 50 Hz and an amplitude that can be adjusted from 0 to 9 kV (using an even higher voltage resulted in electrical breakdown inside the PAP, with sparks bridging the gaps between the copper plates). The dwell time in the plasma zone of the compound studied is approximately 0.6 s. Unless noted otherwise, a voltage of 8.5 kV was used. Using this conditions approximately 300 ppb ozone were measured on the exhaust of the PAP. The humidity during our experiments was determined to be in the 10−14 g/m3 range. Introduction of a sample increased the humidity by no more than 1%. Therefore, a change in the ongoing reactions due to a change in humidity during sample introduction is very unlikely. The emission spectrum of the plasma in the PAP was measured with an optical fiber coupled PC-card-based UV−vis spectrometer (ISS-UV−vis, Ocean Optics, Dunedin FL/USA). For these measurements all ambient light sources were switched off. The background was measured at the same position without activated PAP. The spectrum obtained was background subtracted and shows a number of emission lines (Figure S1 of the Supporting Information). The characteristic line at 779 is due to O emission, and the signals between 300 and 450 nm originate most likely from N2 transitions of the second positive system (C3Πu → B3πg).23 No indications of reactive species were visible in the emission spectrum, presumably because O and N2 emissions are much stronger. The ionization interface (C) consists of a heating coil (1), which was used to heat the aluminum body (3) to 473 K. This greatly helped to reduce memory effects between experiments. The ionization of the PAP exhaust was either accomplished using the PAP itself or the miniaturized atmospheric-pressure afterglow ion source 2 (miniFAPA). The settings for the miniFAPA ion source were 0.50 kV, 15.0 mA, and a gentle helium (∼100 mL min−1, 99.999%; PanGas, Dagmersellen, Switzerland). The ionization processes of the miniFAPA ion source are similar to
Unfortunately, it is not at all straightforward to investigate which products are generated in the plasma because the selection of an adequate sorption material is very problematic for unknown products.20 Moreover, in a worst-case scenario, nontoxic compounds could be modified in such way that undesired toxic products are formed upon passing the PAP. Here, a commercially available PAP was studied using a newly designed interface that allows us to directly lead the exhaust of the PAP into a mass spectrometer (MS). This interface helps to circumvent passive sampling and simplifies the understanding of the molecular processes in the air purification system. The ionization of the compounds in the PAP exhaust was either accomplished using the plasma air purifier itself or with a miniaturized atmospheric-pressure afterglow ion source (miniFAPA).21 To study the performance of the newly designed MS interface, volatile amines were chosen as reactants since it is well-known that amines are easily ionizable. By choosing well-defined starting compounds, we also aimed at reducing the complexity of the product distribution for a comprehensive characterization of the molecular processes in the PAP.
■
EXPERIMENTAL SECTION Materials and Sample Preparation. Water was purified in-house using a Millipore system (Bedford, MA, USA). Methanol, ethanol, as well as hexyl-, octyl-, decyl-, tert-butyl-, dipentyl-, dibutyl-, and d18-dibutylamines (HPLC grade) were purchased from Sigma Aldrich (Buchs, Switzerland). Solutions of 1% by weight of all amines in water, water:methanol (1:1, v:v) or ethanol were prepared. Using an air flow of 320 L min−1 through the PAP, and an ionfusion rate of 40 μL min−1, this a concentration of roughly 1 ppmV results in the ventilation system for all amines. Dilutions were done with the methanol/ water mixture unless noted otherwise. Experimental Setup. The schematic of our experimental setup is shown in Figure 1. It consists of four parts. The sample introduction into the system (A) was accomplished using a syringe pump (Fusion 400, KR Analytical Ldt, Cheshire, UK). The syringe pump allows delivering an exact amount of sample per time, which renders comparisons between experiments very reliable. The sample is nebulized using a pneumatically assisted supersonic-like source.22 For this, a high voltage is needed. Air was used as nebulization gas (1.0 L min−1). The flow through 4068
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
discharge plasma. Upon injecting dipentylamine, the intensity observed of the parent ion (m/z = 158) was more than 1 order of magnitude higher using the miniFAPA ion source compared to the PAP. In the mass spectra obtained using the PAP, a huge signal could be observed at m/z = 378. This signal must be generated in the PAP when dipentylamine is injected, since it was not observed with the miniFAPA ion source. This also shows that the compound causing this peak is not present in solution. The two signals at m/z = 174 and m/z = 190 represent oxidation products of the alkyl chain. Such products have been described in previous work conducted with common volatile organic compounds.20 The same oxidation products could also be detected using the miniFAPA ion source. The signal-to-noise ratio when applying the miniFAPA was not as good as for the PAP, tough. Finding signals with m/z higher than 158 (except for the oxidation products) was surprising, as degradation products were expected to be formed exclusively. In a next set of experiments, the mass-to-charge range was increased to 2000 and the dependence of the voltage applied to the PAP was studied using different amines. Figure 3 shows spectra for octylamine and dibutylamine (parent ions = 130 m/z) without using the miniFAPA ion source. The voltage of the PAP was gradually increased from 6.0 to 9.0 kV in steps of 0.5 kV. For all amines measured, oligomer-like signal series were observed in the mass spectra from 7.0 kV on upward, especially in the case of primary amines. The results obtained show that the higher the applied voltage, the higher the degree of oligomerization. No decomposition products were found. Using the highest possible voltage of 9.0 kV, the parent ion of octylamine (m/z = 130) was much less intense compared to the most intense product ions. It seemed that a large fraction of octylamine introduced reacted to higher molecular weight oligomers. When studying dipentylamine (parent ion = 158 m/z) at an applied voltage of 7.0 kV or higher, the most intense signal changed from the parent ion (m/z = 158) to the first oligomerization product (m/z = 378). This shift toward higher molecular weight products when applying a high voltage to the PAP was found for all compounds tested. The mass steps of 14 and 16 Da observed are very likely due to OH radical side chain oxidations forming alcohols or ketones/aldehydes.1,24−26 Kalberer et al. have previously shown that reactions of carbonyls and their hydrates can generate polymers, which were found to be a substantial fraction of the organics on aerosols. Similar results, with regular mass differences of m/z = 14, 16, and 18 were observed in our study. By increasing the voltage, the maximum of the polymer mass shifted gradually to higher masses. Kalberer et al. observed analogous effects when they increased the reaction time. Figure 4 shows details of four different amines studied. For dipentylamine (parent ion = 158 m/z), a common oligomerization pattern with steps of m/z = 220 was observed. We first assumed that dipentylamine was strongly oxidized to a reactive oxygen radical intermediate on the copper plates with an overall molecular weight of 221 Da that further reacted with ammonium ions passing the PAP. A reactive intermediate with a mass of 221 Da was assumed since a covalent bond formation between the dipentylammonium ion and the reactive intermediate should lead to a loss of 1 Da. The mass difference between the putative reactive intermediate and the dipentylamine is 64 Da. For tert-octylamine, octylamine and hexylamine the very same mass difference between the reactand and the intermedtiate were observed.
those taking place in common atmospheric pressure chemical ionization sources, with active species of the composition [H2O]nH+.20 The vacuum port 4 allowed to aspirate 10% (∼33 L min−1) of the PAP exhaust through the ionization interface, which was accomplished by connecting it to a vacuum pump (DIVAC 2.2 L, Oerlikon, Cologne, Germany). The tip of the elongated MS cone 5 was placed within the aluminum body at a distance of 1.0 cm from the miniFAPA ion source. Mass spectrometric analysis was performed using an ion trap instrument (D) (LCQ Deca, Thermo Finnigan, San Jose, USA) equipped with a self-made inlet cone. The settings of the LCQ instrument were: capillary voltage, 30 V; capillary temperature, 473 K; tube lens, 55 V. The ion optics and main RF were optimized to obtain the best overall ion yield. The LCQ instrument was controlled by the Xcalibur software, version 2.07 SP1.
■
RESULTS AND DISCUSSION In a first set of experiments, the performance of the newly designed interface was studied. It was already known that the plasma air purifier itself was capable of ionizing substances that were present in the air ventilation system. To compare this observation to a better known ionization method, we incorporated the miniFAPA ion source into the interface. Forty microliters of a 1% dipentylamine solution in water was sprayed for 1 min into the system. The concentration transferred into the interface was calculated to be ∼1 ppmV. Figure 2 shows the
Figure 2. Performance test of the newly designed interface injection dipentylamine for 1 min. The upper graph show mass spectrum using the miniFAPA ion source for ionization (PAP turned off). The lower graph show the same, but using only the PAP as ionization source.
mass spectra using either the miniFAPA ion source (upper panel, PAP turned off) or the PAP (lower panel) for ionization. Helium was used in the miniFAPA to generate an active species, which reacts with the analyte. The ionization process resembles the one observed in most plasma ion sources presumably involving mostly charged water clusters ([H2O]n H+). No details are known about the ionization mechanisms in the PAP, but we assume that it is similar to that in a corona 4069
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
Figure 3. Illustration of the influence of the voltage applied to the PAP von the mass spectra when injecting octyl- and dipentylamine. The voltage was gradually increased in 0.5 kV steps from 6 to 9 kV. Every spectrum was obtained by summing up 1 min of data. The m/z range was set to 50−2000.
nitrate, the exhaust of the PAP was bubbled through water for 4 h. Ion chromatography measurements of the water supported the production of nitrate and nitrite (NOx) by the activated PAP. In a control experiment without the PAP activated, only traces of nitrate but no nitrite could be detected. The signal at m/z = 254 could thus be assigned to a cluster of protonated dibutylamine and two nitrate anions. Ionic cluster formation with nitrate explains the observed mass steps shown in Figure 4, and it is in an excellent agreement with the observed masses. Part C of Figure 5 shows the positive mode mass spectrum, along with the suggested sandwich clusters between the positively charged dibutlyamine ions and the nitrate anions, which were obtained when injecting dibutylamine into the PAP. The higher the voltage of the PAP, the more nitrate was generated and the larger and therefore heavier the ionic clusters became. Primary ammonium molecules can form more clusters because steric hindrance could at least partially explain the much more crowded mass spectra observed for primary compared to secondary amines. MS/MS experiments showed that the high mass ionic clusters are related to the lower mass ionic clusters. Isolating the signal at m/z = 706 generated from dipentylamine and using an arbitrary collision energy of 30% in the MS/MS mode, a product ion of m/z = 514 was observed. Using MS3 (1, 706; 2, 514) it was possible to detect a signal at m/z = 322 and even dibutlyammonium could be seen conducting MS4. Therefore, it
In a next set of experiments, we tried to verify the hypothesis of covalent bond formation and therefore prove the existence of a reactive intermediate which would explain the results observed above. Dibutylamine (parent ion = 130 m/z) was chosen as model compound because the mass spectra were not as complex as those for the primary amines. For this reason, per deuterated dibutlyamine (parent ion = 148 m/z) was studied in the PAP. Part A of Figure 5 shows mass steps of m/z = 15 observed for the side chain oxidation products, which is 1 Da less than in the experiments with undeuterated dibutylamine. This can only be explained by the loss of deuterium instead of hydrogen. For two covalent bonds being formed and the consequential loss of two deuterium atoms, oligomerization steps of Δm = 208 instead of 210 would be expected. However, a Δm = 210 was still observed. Therefore, the formation of a reactive intermediate that forms covalent bonds had to be excluded to explain the observed oligomerization series in the mass spectra. Repeating the experiments using water, methanol and ethanol as solvents had no discernible influence to the mass spectra observed. When running the LCQ-MS in negative ion mode, a signal at m/z = 62 was found. Part B of Figure 5 shows the mass spectrum of dibutylaime in negative ion mode. In addition to the negatively charged ion at m/z = 62, an intense signal at m/z = 254 was detected. The signal observed at m/z = 62 could correspond to nitrate, which is known to form from N2 in an air-operated corona plasma.27−29 To prove the existence of 4070
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
Figure 4. Six mass spectra of different amines studied are presented. In each of those spectra, regular mass steps could be observed: regular mass steps of 220 were found for dipentylamine, 192 for dibutylamine, 164 for hexylamine, 192 for tert-octylamine, and 220 for decylamine.
Complete breakdown of the amines studied, as stipulated by manufacturers of such devices, was not observed. On the contrary, higher molecular weight products are generated. It is quite possible that even higher mass polymers are generated that are not detected in the MS, since the instrument used here was optimized for the mass range m/z = 100−1000. Signal intensity in the higher mass range is lost, that is, the intensity of even higher molecular weight products, if present, are discriminated against. We assume the reactions in the PAP to similar in a fashion similar to those proposed by Atkinson for the tropospheric gas phase chemistry30 (abstraction of an H atom by OH radicals; O2 addition to the alkyl radical, forming an alkyl peroxy radical; further reaction with NO to the alkyl radical and with NO2 in an isomerization reaction to a hydroxyl alkyl radical). However, pinning down details of the reaction mechanisms that occur in the PAP has to await further, detailed study. It is not known if the compounds formed in the PAP are toxic or not. Therefore, it is generally recommended to study other air purifiers based on plasma using an approach similar to the one shown in this work. Our results suggest that the design of the PAP studied in this work needs to be improved. In its
is clear that the ionic clusters observed were from the dibutlyammonium interacting with nitrate, which is generated in the PAP. Figure 6 shows the MS/MS of the first oxidized ionic cluster. Both signals at m/z = 130 and at m/z = 146 can be seen, showing that the oxidized cluster of the oxidized dibutlyammonium, the dibutlyammonium and the nitrate is formed. This study shows that most products in the PAP are generated from reactions with OH radicals and subsequent formation of ionic clusters involving nitrate. The distance from the exhaust of the PAP to the LCQ-MS is more than 1 m, that is, the ionic generated clusters are rather stable. The method of coupling the PAP directly to the MS offers significant advantages compared to passive sampling methods (e.g., using adsorption tubes): (i) a short measurement time, which allows covering a wide range of experimental parameters; (ii) first results for compounds of interest are obtained within minutes; (iii) adjusting the setup to find the most efficient experimental conditions is easily possible; (iv) neither timeconsuming desolvation nor sample preparation steps are necessary; (v) the compounds to analyze do not need to be known in advance to get valid results (no need to select different adsorbing media). 4071
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
Figure 6. (Top) Mass spectrum of the product distribution generated from dibutylamine after passing through the PAP, (bottom) MS/MS spectrum of the first oxidized ionic cluster at m/z = 338. Collision energy was kept at 30% (arbitrary unit).
A special vent was mounted over the (ambient pressure) inlet to the mass spectrometer to guarantee safe working conditions.
■
ASSOCIATED CONTENT
S Supporting Information *
Figure 5. (A) Mass spectrum of the deuterium labeled dibutlyamine studied in the PAP. Regular mass steps of 15 and 210 are found. The mass steps of 15 confirm clearly the side chain oxidations. (B) Mass spectrum of dibutlyamine measured in the negative ion mode. Nitrate and a nitrate-dibutlyammonium-nitrate ionic cluster could be seen. (C) Mass spectrum and schematic representation of the polymeric products generated from dibutylamine in the PAP.
Chart of intensity vs wavelength. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected], Tel: +41 44 632 4376. Notes
current state, it cannot be recommended as an all-round air purification system because it is not known what influence the molecules generated have on public health. The production of high-mass polymers from small compounds could also present an alternative strategy for removing volatile compounds from polluted air because polymeric compounds are nonvolatile. Nonvolatile species could easily be trapped on an electrostatic filter.
■
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Christoph Bärtschi from the mechanical shop is gratefully acknowledged for manufacturing the MS interface as well as the whole experimental setup. We also thank Reto Glaus for assistance using the ion chromatography system.
■
SAFETY CONSIDERATIONS
REFERENCES
(1) Atkinson, R.; Carter, W. P. L Kinetics and mechanisms of the gasphase reactions of ozone with organic compounds under atmospheric conditions. Chem. Rev. 1984, 84 (5), 437−470.
We made sure that the entire gas exhaust of the used experimental setup was guided into the laboratory gas exhaust. 4072
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073
Environmental Science & Technology
Article
branched, and cyclic alkanes in the presence of NO(x). Aerosol Sci. Technol. 2009, 43 (6), 604−619. (26) Aschmann, S. M.; Arey, J.; Atkinson, R. Atmospheric chemistry of three C-10 alkanes. J. Phys. Chem. A 2001, 105 (32), 7598−7606. (27) Burlica, R.; Kirkpatrick, M. J.; Locke, B. R. Formation of reactive species in gliding arc discharges with liquid water. J. Electrost. 2006, 64 (1), 35−43. (28) Huang, L. P.; Yang, S.; Liu, S. J.; Wang, Q. H.; Zhu, Y. Z. Removal of NO(2) produced by corona discharge in indoor air cleaning. J. Adv. Oxid. Technol. 2009, 12 (2), 238−241. (29) Atkinson, R. Atmospheric chemistry of VOCs and NOx. Atmos. Environ. 2000, 34 (12−14), 2063−2101. (30) Atkinson, R. Gas-phase tropospheric chemistry of volatile organic compounds 0.1. Alkanes and alkenes. J. Phys. Chem. Ref. Data 1997, 26 (2), 215−290.
(2) Hu, G. P.; Ran, P. X. Indoor air pollution as a lung health hazard: Focus on populous countries. Curr. Opin. Pulmonary Med. 2009, 15 (2), 158−164. (3) Jones, A. P. Indoor air quality and health. Atmos. Environ. 1999, 33 (28), 4535−4564. (4) Zuskin, E.; Schachter, E.; Mustajbegovic, J.; Pucarin-Cvetkovic, J.; Doko-Jelinic, J.; Mucic-Pucic, B. Indoor air pollution and effects on human health. Periodicum Biologorum 2009, 111 (1), 37−40. (5) Ingrosso, G. Free radical chemistry and its concern with indoor air quality: An open problem. Microchem. J. 2002, 73 (1−2), 221−236. (6) Sarwar, G.; Corsi, R.; Kimura, Y.; Allen, D.; Weschler, C. J. Hydroxyl radicals in indoor environments. Atmos. Environ. 2002, 36 (24), 3973−3988. (7) Mentese, S.; Arisoy, M.; Rad, A. Y.; Gullu, G. Bacteria and fungi levels in various indoor and outdoor environments in Ankara, Turkey. Clean Soil Air Water 2009, 37 (6), 487−493. (8) Hall, J. V.; Winer, A. M.; Kleinman, M. T.; Lurmann, F. W.; Brajer, V.; Colome, S. D. Valuing the health benefits of clean air. Science 1992, 255 (5046), 812−817. (9) Sessa, R.; Di Pietro, M.; Schiavoni, G.; Santino, I.; Altieri, A.; Pinelli, S.; Del Piano, M. Microbiological indoor air quality in healthy buildings. Microbiologica 2002, 25 (1), 51−56. (10) Spengler, J. D.; Sexton, K. Indoor air pollution. A public-health perspective. Science 1983, 221 (4605), 9−17. (11) Lazaridis, M. Indoor Air Pollution. Springer: p 255−304. (12) Holzer, F.; Roland, U.; Kopinke, F. D. Combination of nonthermal plasma and heterogeneous catalysis for oxidation of volatile organic compounds Part 1. Accessibility of the intra-particle volume. Appl. Catal., B 2002, 38 (3), 163−181. (13) Kohno, H.; Berezin, A. A.; Chang, J. S.; Tamura, M.; Yamamoto, T.; Shibuya, A.; Hondo, S. Destruction of volatile organic compounds used in a semiconductor industry by a capillary tube discharge reactor. IEEE Trans. Ind. Appl. 1998, 34 (5), 953−966. (14) Roland, U.; Holzer, F.; Kopinke, F. D. Improved oxidation of air pollutants in a non-thermal plasma. Catal. Today 2002, 73, 315−323. (15) Subrahmanyam, C.; Renken, A.; Kiwi-Minsker, L. Novel catalytic non-thermal plasma reactor for the abatement of VOCs. Chem. Eng. J. 2007, 134 (1−3), 78−83. (16) Subrahmanyam, C.; Renken, A.; Kiwi-Minsker, L. Novel catalytic dielectric barrier discharge reactor for gas-phase abatement of isopropanol. Plasma Chem. Plasma Process. 2007, 27 (1), 13−22. (17) Van Durme, J.; Dewulf, J.; Sysmans, W.; Leys, C.; Van Langenhove, H. Abatement and degradation pathways of toluene in indoor air by positive corona discharge. Chemosphere 2007, 68 (10), 1821−1829. (18) Gao, D.; Yang, X. C.; Zhou, F.; Wu, Y. A. Experimental study on indoor air cleaning technique of nano-titania catalysis under plasma discharge. Plasma Sci. Technol. 2008, 10 (2), 216−220. (19) Eliasson, B.; Kogelschatz, U. Nonequilibrium volume plasma chemical processing. IEEE Trans. Plasma Sci. 1991, 19 (6), 1063− 1077. (20) Schmid, S.; Jecklin, M. C.; Zenobi, R. Degradation of volatile organic compounds in a non-thermal plasma air purifier. Chemosphere 2010, 79 (2), 124−130. (21) Jecklin, M. C.; Schmid, S.; Urban, P. L.; Amantonico, A.; Zenobi, R., Miniature flowing atmospheric-pressure afterglow ion source for facile interfacing of CE with MS. Electrophoresis 31, (21), 3597-3605. (22) Hirabayashi, A.; Sakairi, M.; Koizumi, H. Sonic spray ionization method for atmospheric-pressure ionization mass-spectrometry. Anal. Chem. 1994, 66 (24), 4557−4559. (23) Harshbarger, W. R.; Porter, R. A.; Miller, T. A.; Norton, P. Study of optical emission from an RF plasma during semiconductor etching. Appl. Spectrosc. 1977, 31 (3), 201−207. (24) Heimann, G.; Warneck, P. OH radical induced oxidation of 2,3dimethylbutane in air. J. Phys. Chem. 1992, 96 (21), 8403−8409. (25) Lim, Y. B.; Ziemann, P. J. Chemistry of secondary organic aerosol formation from OH radical-initiated reactions of linear, 4073
dx.doi.org/10.1021/es2042492 | Environ. Sci. Technol. 2012, 46, 4067−4073